MAS pilots cross-bank AI models to detect scam risks

All sensitive data used in the exercise will be permanently deleted once complete.

The Monetary Authority of Singapore (MAS) is working with banks, the Government Technology Agency of Singapore, and the Singapore Police Force to strengthen scam detection using artificial intelligence and machine learning (AI/ML).

In a statement, the authority said it is conducting a proof-of-value (POV) exercise to test AI/ML techniques for pre-emptive scam detection, combining data from five banks.

MAS has set up a secure data-sharing environment governed by data protection policies and protocols to support the exercise.

Data used in the POV will remain confidential and protected using cryptographic techniques, whilst bank account numbers will be hashed so that only the originating bank can identify them.

Data access will also be restricted to authorised personnel in a controlled setting, which will be continuously monitored throughout the exercise.

All data used in the POV will be deleted at the end of the exercise.

MAS said the initiative is intended to support industry collaboration in using AI/ML to complement existing anti-scam measures.

Furthermore, the regulator may expand the models by incorporating broader datasets and additional use cases after assessing their effectiveness.